Eric Hanushek on Educational Quality and Economic Growth
Aug 6 2007

Hoover Institution Senior Fellow Eric Hanushek talks about his research on the impact of educational quality on economic growth. Past efforts to increase the economic growth rate of poor countries have focused on years of schooling, neglecting the quality and true education that needs to take place. Hanushek presents dramatic findings about the decisive nature of cognitive ability and knowledge in driving economic growth. Join us as Hanushek talks with EconTalk host Russ Roberts about his findings and the implications for public policy around the world and in the United States.

Eric Hanushek on the Education, Skills, and the Millennium Development Goals
How important are basic skills for economic success and growth? Eric Hanushek of Stanford University's Hoover Institution talks with EconTalk host Russ Roberts about the importance of basic education in math and literacy and their relationship to economic growth. Hanushek...
Lant Pritchett on Education in Poor Countries
Lant Pritchett of Harvard University and author of The Rebirth of Education talks with EconTalk host Russ Roberts about the ideas in the book. Pritchett argues that increases in years of schooling for students in poor countries do not translate...
Explore audio transcript, further reading that will help you delve deeper into this week’s episode, and vigorous conversations in the form of our comments section below.


Aug 6 2007 at 7:01pm

I am a big fan of econ talk and I think that you Russ Robert are a great host.

On this pod cast:
First I think that it is crucial that we explicitly say that schooling and education are not the same thing. IMO Schooling is only a small subset of education. IMO education is crucial but it is hard for me to believe that schooling has much impact on economic performance. My grandparents went to one year of school each and they ran a successful barber shop and to talk to them they seemed more educated than many of todays college graduates (they read). Further how about Thomas Edison, Ben Franklin and Micheal Farrade? Further intuitively why would business not train people if they could get so much more production out of them and therefore economic growth. Further why would parents not see to it that their children not educated accept that they need the money first thus growth precedes schooling and education. IMO causation runs that way.

Further if you look at the work of Richard Vedder it looks like more government spending on education tends to undermine growth.

BTW I think that Richard Vedder would make a great guest.

Aug 6 2007 at 7:34pm

BTW a couple of ideas

1. An idea for a researcher: We are fortunate in this country to have a set of people who are lavished with schooling even to the point of having private tutors starting in high school. Those are the good athletes. Good athletes get tutors some from the high school level. IMO this would be a good set of people to look at for research.

2. I think that our schools are setup more to grade people than to teach and because of this they teach the wrong things. It seems that purpose of early schooling is to prepare people for further schooling. They never seem to get to teaching people what they need to know to do well in life.

Aug 8 2007 at 1:20am

One thing I thought was missed from this, and would like some feedback on this.

When I was getting my Econ degree 15 years ago or so, my college (a commuter type school in the LA area) had a influx of a number of older students who were defense related engineers laid-off as part of the “peace dividend” myth peddled back then. Apparently the labor market for engineers was so tight in the 80’s that recruiters would go on campuses and recruit students without a degree.

Fast forward to today, where I find myself living and working in South Korea. Here they place a particular importance on things like how long you have been in school, and particularly getting a university degree of some sort. All this looks good in numbers, however many of the degrees have limited, or no, recognition outside of Korea.

These things have always made me wonder how effective measurements like “years” or “degrees” are. Or more to the point, how much of a function does a successful economy (i.e. one with a tight labor market) actually has a negative effect on years of schooling?

Larry Willmore
Aug 8 2007 at 5:05am

I enjoyed this podcast, but have one complaint.

Professor Hanushek refers to the effects of ‘decentralisation’ on Swedish schools. This seems a curious way to refer to the opening of Swedish basic education to competition in 1992. Here are some quotes on this subject from a recent paper by A. Böhlmark and M. Lindahl, “The Impact of School Choice on Pupil Achievement, Segregation and Costs: Swedish Evidence”, IZA DP 2786 (May 2007):

Sweden … went from a situation where pupils were assigned to their closest public school (närhetsprincipen), where the possibility of choosing another school was very limited, to a system that allowed pupils to freely choose among both public and private schools. [p. 3]

To be eligible for public funding, private schools … have to follow the national curriculum and are not allowed to choose their pupil bodies. …. Private schools are not allowed to charge any fees. Nor are there any restrictions on the ownership structure of the private schools eligible for public funding — whether religious, non-profit parent cooperatives, or for-profit corporations. [p. 7]

[T]he total achievement effect is mainly driven by other peoples’ choice of private school in the municipality. We interpret this as evidence of competition effects where more school competition forces all schools to improve. [p. 41]

See also the 2005 Journal of Public Economics paper by Sandström & Bergström cited by Hanushek & Wößmann in their February 2007 World Bank paper.

Brad Hutchings
Aug 8 2007 at 7:03pm

Russ, I am glad you pressed him on accountability at the end of the podcast. It’s funny, but my Dad ran for local school board in 1988 on a platform of accountability — before it was cool. Being mostly small-l libertarian in thought (but “Conservative Republican” in name), he didn’t fit into the standard mold of either the teacher union puppet or religious fanatic that you sorta had to fit into to gain a vocal constituency and win a seat.

I am quite troubled in general by the federally led approach to education. To me, NCLB looks like a disaster based on a plan that could lead to nothing else. His comment that you can’t do economics without reading says it all for me. It says that you drag the high end achievers down to ensure that all achieve a minimum. I even noticed this 20 years ago when I was in school. OK, I got why I should take the state competency tests, which were pathetically easy. I didn’t get why I and most of my smarter than average classmates had to spend class time preparing for them. Seriously… Offer a “remedial preparation” elective for those who might need it to pass the test. Let the rest of us spend the time taking art or computers or a class on the stock market ;-).

He rails against individual state standards. Back in the day when I was in elementary school, the California standard drove the nation. It was the benchmark. The book publishers published to it. You didn’t need standards written at the federal level because other states actually adopted what one large state did. Seems to me like kind of a division of labor slash comparative advantage thing among state bureaucracies. Rather than defer to the feds, other states ought to scrape out areas of expertise. Book publishers are a tried and true mechanism for disseminating and sharing the high standards.

And the whole funny thing about testing is that if we can’t solve it in engineering/software domains, how is anyone optimistic that we can solve it in educational domains? My ugrad and grad education involved many software courses where testing was addressed. One professor was involved in reviewing a testing plan for TCAS, which was an automated collision avoidance system for airplanes moving 600mph. And the stories on that one would make you never want to fly or have airplanes pass over your house! That’s 15 years ago. Today, we have whole development methodologies that are test-centric (like “agile” development). But even using these, we still produce buggy, failure-prone, potentially deadly software! The reason is that we generalize about global quality from results of a specific local testing regime, and the leap of logic can be way too much. As Arnold Kling is fond of asking climate scientists to share their modeling science with economists, I am fond of asking educators to share their testing science with software people. Because we really, really, really need the help if they’ve got this all figured out. ;-).

Aug 8 2007 at 11:32pm


I did not follow the last part “But today every state has its own standards: Peach State Standard, Sunshine State Standards. But the problems in algebra don’t change across states. If sequence of learning math matters, the best sequence should drive out the worst.”

There is a huge ideological divide on how to teach math in K-12 and the battleground are the state standards. See this:

I unfortunately don’t see any market forces at play here and I can’t imagine the federal govt stepping in as a solomonic judge.

Also about home-schooling: I think I’ll take advantage of division of labor, even though I might be better at both my job and educating kids.

PS: You mentioned briefly social security reform in Chile. I had heard about SS reform in Sweden but not in Chile, can you give us a link?

Henry Cate
Aug 9 2007 at 12:38am


Have a hard time in the U.S. even estimating how many kids are home schooled. Disappear from system. Estimates of 1%-2%–that’s a 100% difference!

The 1% to 2% number is a five or ten years old. People frequently toss around a 2% to 3% range now. This is based on a survey by the federal government. They sent out about 10,000 surveys. Half of them were returned, around 5,000. Of those about 125 said they were homeschoolers. But there were still 5,000 people who didn’t respond.

There is one school of thought that homeschoolers by their nature are non conformist and are much less likely to comply with requested surveys from the government.

The bottom line is we just don’t know how many children are homeschooled. There is good reason to believe the number is growing between 7% to 10% a year.

karl smith
Aug 9 2007 at 8:08pm

One issue I am interested is whether IQ measures or PISA type measures dominate in growth regressions.

I have seen some evidence that subject test scores dominate the SAT when looking at college performance.

Are intellegence tests proxying for math and reading skill or is it the other way around. Or perhaps they both contribute. I think this is an important point as we think about how to model the cognitive skills => growth connection.

Aug 10 2007 at 6:17pm

As a Middle School teacher in a private school in Mexico, allow me to offer the following for your consideration:

1. Under Mexican regulations, a student may not be failed. Our official grading system goes from 5 to 10, 5 being the lowest possible grade you can give. If you end the year with a 5 average in a subject, you don’t fail but instead must take an “extraordinary exam”. If you get a five on this exam, you can take it again, and then again.

2. The public education secretariat (SEP) has jurisdiction over every single school in the country, both public and private. (The only exception is the American School Foundation which, at over 100 years of age, is older than the secretariat. The ASF pretty much does what it wants.) The SEP requires every school, public and private, to send it, to give just one example, its annual list of field trips along with parent permission slips. It actually pays people to go over them.

3. The teachers union is perhaps the most powerful union in the country and can literally decide a presidential election. Its members are among the best paid among OECD countries (maybe even THE best paid) and yet they also have one of the highest rates of striking. Recently, in the state of Oaxaca, the teachers (for the entire state) were on strike for months, maybe even for most of the year.

Ken Willis
Aug 11 2007 at 9:18am

I perked up when you asked him about home schooling. I wanted to know what he thought of it or what his research showed. I don’t think he ever answered the question.

Aug 13 2007 at 12:26pm

I was disappointed that Hanushek never got specific about what we do know about what does and doesn’t work. Granted a lot is unknown, but it sounded as though he was going to get specific early in the interview when the topic veered off. (Russ, your very good but sometimes you talk when it’s the interviewee we want to hear…)

Hanushek is undoubtedly right that “accountability” is key. But the most effective accountability is that imposed by the market. Hence, by far the most effective reform would be a system that introduced competition, the simplest of which are surely vouchers. Of course, it’s dangerous for politicians and even researchers to say favorable things about vouchers because like Mexico (as I learned from the above comment), the U.S. teachers union is very powerful and very quick to wield that power against enemies via its political clout. In fact, the teacher’s union is the elephant in the room–the main obstacle to reform.

Sep 2 2007 at 1:49pm

I enjoyed the econtalk a lot. I just checked the PISA website and found out that most of the countries with top score do not have strong private education section. This contradictes to the econcomist’s belief that privatizing eduction will help the quality of education. I grew up in HongKong which has the highest score in science and math. However the best schools in Hongkong are public schools and teachers’ performance has nothing to do with their compensations. I wonder if there are other factors affecting the test score.

Sep 14 2007 at 10:36am

Unless I missed something, I find that this discussion from the guest missed out on a fundamental issue: parents. To discuss any of this school subject without examining the critical and central variables involving the qualities of the family’s relationships and intergenerational influences on life at home, is a waste of time.

Comments are closed.


About this week's guest:

About ideas and people mentioned in this podcast:Articles:

Web Pages:

Podcasts and Blogs:



Podcast Episode Highlights
0:36Intro. How does education affect growth and development? What have we learned about the relationship between education and growth? What determines growth across nations? Human capital, but it's hard to measure. School attainment--number of years of schooling--has been the standard measure but it's not a very good measure of human capital, particularly in an international context. Gary Becker, early 1960s, Ted Schultz, Adam Smith, skill of the population. Stock of knowledge, know-how, embedded in people; you can invest in it; it depreciates. Years of schooling is positively related to differences in growth rates, but what if the explanation is just that wealthier nations just buy more schooling because they buy more of everything? No one thinks that a year of fourth grade education in Egypt is the same a year of fourth grade education in the U.S. But studies using attainment treat a year as a year as a year. Relating test scores to growth is a much more precise measure. U.S. debate often is about how some kids in grade four know more than others in grade four and how can we make them the same? That's multiplied in spades in international comparisons. South America has typically had a lot of schooling, but putting them on the same scale as the U.S. by comparing math and science by grade nine turns out to be abysmal. Different backgrounds in the home, different IQs: Is it different schools or different backgrounds? Parents have a lot to do with education. Health and nutrition makes a difference. Is school quality an important determinant of cognitive abilities? Yes--different schools, different teachers can get a lot more learning out of kids. Similar internationally. But family background also matters in both the U.S. and internationally.
8:14Anecdotes: we think some teachers affected us and we remember it. Is that controversial? It's not controversial that teacher quality matters. What is controversial is "What do you make of that?" How do you affect the quality of teachers, is controversial. Differential pay, performance pay: how do you measure performance, attract better quality teachers? Principals, parents, and other teachers know, within a range. Who are the very best and very worst is known, but the middle is harder. Objective measures of teacher quality as the basis of compensation is difficult. Subjective measures are used all the time in business, which is not a public arena. Unions typically don't want to make these distinctions. What is value added of the teacher separated out from the classroom, etc.? In business, managers are rewarded directly for their performance, so they have an incentive not to randomly distribute funds to those they hire. In schools we don't have performance pay for principals and superintendents, so it's hard to have performance pay for the teachers they hire. Private schools: principals are not compensated on student performance, but they can be fired by the parents. But why isn't there more merit pay in the private sector? Many private schools don't volunteer that data, though. Looks like there is more pay variation in private than public schools. Not the variation in pay so much as judgments and retention decisions. Ability to fire and hire, blunter stick than pay changes.
15:20International sector. Cognitive ability is correlated with growth rates. Since 1960s series of international tests, math, science, recently in reading, PISA test, Programme on International Student Assessment, OECD (Organization of Economic Cooperation and Development. In Germany these tests became the subject of intense national discussion, shocked that with these tests their school system didn't come out the best in the world. Tests begun in 2000. Since the mid-1960s there have been 15 or so different tests given internationally; historically U.S. has been in the middle or below average, particularly when you talk about the high school level. Doesn't U.S. represent a wider population? Yes, but not so relevant today as in earlier tests. Looking at developing countries find that they are nowhere near the U.S. or Europeans. In poor countries, even those who can afford to stay in school are not getting as good an education. In Peru, only 8% of the kids taking the test are above the bottom 15% of the European kids. Typical of developing countries. World Bank has extensive education programs in developing countries: "Education for All" and "Lending and Development Goals". Want by year 2015 to get all children should get to junior high level (lower secondary level). Focus is on years of schooling neglects issue of whether they are learning anything while in school. Evidence from cognitive tests suggests that it's what they know that counts for economic growth, not how long they've been in school. Have to focus on quality of schooling, which is something that can potentially be influenced governmentally. World Bank program works by making loans at low interest rates, sometimes no interest, for social programs, bargaining situation. Buying new buildings, however helpful, may not change the teachers doing the educating in the schools. Attendance by both students and teachers can be a problem in these countries. Low teacher pay can mean teachers have second jobs.
27:11Growth rate correlation: If you could move the average kid in Mexico up to the level of the average kid in the U.S. in math and science, the Mexican growth rate could improve by as much as 2% per year. Large when compounded over time. If you could move the U.S. up to the level of a good European country (Denmark, Netherlands, Sweden), U.S. growth could improve 1/2 to 1% per year because of improved skill of workers. These are long-run effects because it takes a while for kids to get out of school. If we could take a poor country like Peru and starting in pre-school improve the quality of schooling through high school, are you confident in the finding that it would improve growth rates? Very confident. Other factors like security of property rights, open trade, culture, religious background; but quality of schooling factors hold up regardless of specific measures of cognitive skills. Hanushek and co-author Ludger Wößmann (Woessmann) have been looking at: over time some countries get better and others get worse on these tests. Growth rates on countries that have gotten better have jumped up, and for those that get worse have fallen back. Powerful evidence, suggesting it's not just IQ. Finland: dramatic increases in student performance in last 30 years, maybe nothing left to do in Finland, anecdotally reasons not known. Norway became much more focused on school quality but their school quality fell. Sweden decentralized their schooling system and improved quality on tests; but we don't really know the reasons for improvement. Emigration changes results as well; test scores are not separated by natives vs. immigrants. California student excuse for their doing worse on national tests is often attributed to immigration, but California does almost equally as badly for both immigrants and non-immigrants. Averages still low.
37:47What recommendations for organizations like the World Bank or for nations to increase quality of schooling? Incentives to get better teachers. Problem with just doubling the salary of teachers, but "bad teachers like more pay as much good teachers." Maybe even more because they have worse alternatives. Some institutional features make a difference. Better accountability for student outcomes makes a difference. More autonomy at local level coupled with accountability makes a difference. Some evidence that more parental choice on what schools their children go to makes a difference. Chile, Columbia have experimented with choice. Unionization and public provision of schools are common; public provision is associated with less accountability. In U.S. private schools at least at university level but not at primary and secondary level. In the Netherlands large private religious school sector, required to follow same standards as public sector, different from U.S. Hope of radical reform seems to be low. Political process in a country is unlikely to take a radical stab at reform. Any examples of experiments as dramatic as Chile's Social Security reform? Some national and private experimentation. In developing countries some movement toward actually running experiments, allows confidence in answers if experiments are being well done. India, Chile. Does it make a difference if you provide a flip chart in paper to write on or not (no blackboard)? Monitoring if teacher is in attendance. Vouchers in Chile, Columbia. World Bank should be looking at these experiments closely. Bigger impact than buying whole programs. Is World Bank the largest potential reformer in the world? Domestic national countries larger by comparisons. U.N. has "Lending and Development Goals" but is not a funder. Some private foundations supporting the experiments.
46:43Home schooling in the U.S.: is it a potential solution? By opting out does it put pressure on schools to improve? Does it improve class size issues? Any evidence about quality and cognitive ability? None. Have a hard time in the U.S. even estimating how many kids are home schooled. Disappear from system. Estimates of 1%-2%--that's a 100% difference! Large number of kids, but we don't even know how many. Not taking the standardized tests. Accountability means getting information about student performance and making it public. Parents have a hard time figuring out how good their kids' schools are. "No child left behind" has had the effect of parents asking their schools why their kids aren't doing better, aren't doing as well as the kids down the road. Information has to be available on the quality of the schools, both locally and nationally. Developing countries have not traditionally taken part in this. Teaching to the test, though, can be an excuse. Scoring high on a math test is not the only goal. Need to have high-quality tests, in which case there would be no qualms about teaching to the test. Have to identify the things we care most about. Now, math and reading. Don't want schools to do nothing else; but little evidence schools do that. If schools did more math and reading and less economics, is that desirable? Can't do much economics if you can't read. High school economics is often structured as a finance course, stock market class, which is not what economics is about. Hard to devise a good test of analytical ability, critical thinking, creativity. Can have standardized tests that test skills of higher order of thinking. Hard to measure creativity, but maybe you can't teach it. Can't say "they're not teaching math so they must be teaching creativity." Art not taught to most talented students after a certain level, not forced. Accountability is a matter of getting the right balance. Have to be able to do some algebra problems.
56:44In a private system the standards of accountability would emerge from the competitive process. Imposing a national accountability standard does force everybody into the same box. Math, reading, maybe a little science would be that box. Our economy is built on free exchange of goods and labor across the states. Federal Government could provide some leadership on what the standards are across the states. No reason to reinvent the standards in Washington, D.C. versus California. But today every state has its own standards: Peach State Standard, Sunshine State Standards. But the problems in algebra don't change across states. If sequence of learning math matters, the best sequence should drive out the worst. Has web helped the process of information sharing? No, more a matter of the development of the information. We don't have a culture of learning what goes on in the schools--in the U.S. or in the rest of the world. We assume there is a right answer, but we don't learn enough to throw out the bad ideas. In California and in many other states we don't have the information available to say if a program is doing a good job or not. Good for education researchers, though!

More EconTalk Episodes